Longitudinal records supported forecasts of multiple sclerosis progression
Source check: Source checked. Not independent scientific confirmation.
Finding: The study used routine records from 1,624 people with multiple sclerosis to forecast transition from relapsing-remitting to secondary-progressive disease at 180, 360 and 720 days. Visit-based models had 70–100% recall but 5–10% precision without confidence filtering; the history model reached 67% recall and 42% precision at 720 days. Few transitions and false positives limit treatment use; the study does not show clinical benefit.
Technology: Recurrent neural network (LSTM); random forest, SVM, KNN and AdaBoost
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- The models learned from clinical visits or longitudinal patient history to forecast disease-course transition.
- Test and result
- A single-center retrospective cohort of 18,574 visits from 1,624 patients. Visit models used patient-wise leave-one-group-out evaluation; the history model used a train/test split because leave-one-group-out training was computationally infeasible.
- Limitation / still unknown
- No independent, prospective or clinical-impact evaluation is reported unless specified in the tested field.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://doi.org/10.1371/journal.pone.0230219
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Read the original full text, including author affiliations, Methods and Results; checked the first result-bearing version and active/archive catalog for same-result records..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.